Comparative Efficacy of First-Line Treatments of Chronic Lymphocytic Leukemia: Network Meta-Analyses of Survival Curves

Abstract

Background

Multiple treatment options in first-line chronic lymphocytic leukemia (CLL) pose a challenge in identifying the best treatment. We performed novel network meta-analyses (NMA; 8 trials, 11 treatments) on the Kaplan-Meier curves to compare treatments for fludarabine-ineligible patients on progression-free survival (PFS), time-to-next-treatment (TTNT) and overall survival (OS).

Methods

Using the Guyot method of enhanced secondary analysis of digitized survival data and applying the fixed lognormal distribution model, we extracted the survival proportions and hazard ratios (HR) over 60 months of follow-up, including PFS comparisons by unmutated/mutated IGHV and del 17p.

Results

Acalabrutinib-plus-obinutuzumab was associated with higher 5-year PFS proportions than ibrutinib (HR = 0.42, 95% CrI = 0.25–0.63) but not acalabrutinib, ibrutinib-plus-obinutuzumab, ibrutinib-plus-rituximab or venetoclax-plus-obinutuzumab. In patients with un-mutated (but not with mutated) IGHV higher PFS proportions and favorable HRs were observed for acalabrutinib, acalabrutinib-plus-obinutuzumab, and ibrutinib-plus-obinutuzumab relative to ibrutinib; and targeted therapies were superior over chemoimmunotherapies in patients with del 17p. Targeted therapies containing ibrutinib or acalabrutinib regimens were associated with superior TTNT over venetoclax-plus-obinutuzumab and all chemoimmunotherapies. OS NMA generally found no difference between therapies except for some chemoimmunotherapies.

Conclusions

Overall, only acalabrutinib-plus-obinutuzumab was associated with superior 5-year PFS gains over ibrutinib, which in turn was similar or superior in PFS benefit over other targeted therapies. Acalabrutinib and ibrutinib with obinutuzumab and acalabrutinib monotherapy were associated with greater 5-year TTNT benefits. Despite marked 5-year OS for many regimens, a differential 5-year OS benefit could not be ascertained.